(361f) A Study On Steam Reforming Of Methanol In A Pd-Based Membrane Reactor Fabricated By Surfactant Induced Electroless Plating

Authors: 
Islam, M. A. - Presenter, North Carolina A&T State University
Akanda, M. H. - Presenter, North Carolina A&T State University
Ilias, S. - Presenter, NORTH CAROLINA A&T STATE UNIVERSITY


Steam reforming of methanol was investigated in Pd-based stainless steel microporous membrane reactor. The Pd-composite membrane was fabricated on microporous stainless steel tubular support by novel electroless deposition technique in which surfactants were used to enhance the Pd deposition and its surface morphology. Membrane surface morphology was characterized by SEM (Scanning Electron Microscopy), EDS (Energy Dispersive Spectroscopy), XRD (X-Ray Diffraction) and AFM (Atomic Force Microscopy). Single and mixed gases were used to measure the H2-perm-selectivity. The membrane-reactor was configured in shell-tube structure. The reaction was carried out on the shell side and the permeate hydrogen was taken out in the tube side. The reaction was carried out over commercially available Cu/ZnO/Al2O3 catalyst. A two-dimensional, pseudo-homogeneous membrane-reactor model was developed to investigate the steam-methanol reforming reactions in a Pd-based membrane reactor. Radial diffusion was taken into consideration to account for the concentration gradient in the radial direction due to hydrogen permeation through the membrane wall. With appropriate reaction rate expressions, a set of partial differential equations was derived using the continuity equation for the reaction system. The equations were solved by finite difference method. The model was also capable of handling steam-methanol reforming reactions under non-membrane condition and equilibrium reaction conversions. Experimental results shows membrane reactors with relatively higher selectivity values, have considerably higher conversion than that of non membrane counter part. Experimental methanol conversions were found to be in good agreement with the model predictions. Some of our experimental and modeling results will be presented in this paper.